Video conference applications are becoming ubiquitous on mobile devices. Existing implementations of decode and encode pipelines on many graphics processing unit (GPU) hardware are for general purpose usage. Decode pipelines are typically capable of supporting media content of different profiles and encode pipelines are typically capable of generating high quality output regardless of the amount of motion in the content.
Typically, content in video conferencing usage does not have a lot of motions when adult users are just chatting. However, sometimes the users can be active children having an increased amount of motion. Likewise, content in video conferencing usage may have a lot of motions when users are perform panning of the surrounding environment.
Hierarchical motion estimation (HME) and Super HME are quick sync. video features that may typically be used to improve quality in encoded content. For example, HME and/or Super HME may improve quality in encoded content when the content has a lot of motions. However HME and/or Super HME may cost extra power because HME and/or Super HME may consume both GPU processing cycles and memory bandwidth.